Amino acid residues that define both the isoprenoid and CAAX preferences of the Saccharomyces cerevisiae protein farnesyltransferase. Creating the perfect farnesyltransferase.

Studies of the yeast protein farnesyltransferase (FTase) have shown that the enzyme preferentially farnesylates proteins ending in CAAX (C = cysteine, A = aliphatic residue, X = cysteine, serine, methionine, alanine) and to a lesser degree CAAL. Furthermore, like the type I protein geranylgeranyltransferase (GGTase-I), FTase can also geranylgeranylate methionine- and leucine-ending substrates both in vitro and in vivo. Substrate overlap of FTase and GGTase I has not been determined to be biologically significant. In this study, specific residues that influence the substrate preferences of FTase have been identified using site-directed mutagenesis. Three of the mutations altered the substrate preferences of the wild type enzyme significantly. The ram1p-74D FTase farnesylated only Ras-CIIS and not Ras-CII(M,L), and it geranylgeranylated all three substrates as well or better than wild type. The ram1p-206DDLF FTase farnesylated Ras-CII(S,M,L) at wild type levels but could no longer geranylgeranylate the Ras-CII(M,L) substrates. The ram1p-351FSKN FTase farnesylated Ras-CIIS and Ras-CIIM but not Ras-CIIL. The ram1p-351FSKN FTase was not capable of geranylgeranylating the Ras-CII(M,L) substrates, giving this mutant the attributes of the dogmatic FTase that only farnesylates non-leucine-ending CAAX substrates and does not geranylgeranylate any substrate. These results suggest that the isoprenoid and protein substrate specificities of FTase are interrelated. The availability of a mutant FTase that lacked substrate overlap with the protein GGTase-I made possible an analysis of the role of substrate overlap in normal cellular processes of yeast, such as mating and growth at elevated temperatures. Our findings suggest that neither farnesylation of leucine-ending CAAX substrates nor geranylgeranylation by the FTase is necessary for these cellular processes.

Mutational analysis of the beta-subunit of yeast geranylgeranyl transferase I.

The gene CAL1 (also known as CDC43) of Saccharomyces cerevisiae encodes the beta subunit of geranylgeranyl transferase I (GGTase I), which modifies several small GTPases. Biochemical analyses of the mutant enzymes encoded by cal1-1, and cdc43-2 to cdc43-7, expressed in bacteria, have shown that all of the mutant enzymes possess reduced activity, and that none shows temperature-sensitive enzymatic activities. Nonetheless, all of the cal1/cdc43 mutants show temperature-sensitive growth phenotypes. Increase in soluble pools of the small GTPases was observed in the yeast mutant cells at the restrictive temperature in vivo, suggesting that the yeast prenylation pathway itself is temperature sensitive. The cal1-1 mutation, located most proximal to the C-terminus of the protein, differs from the other cdc43 mutations in several respects. An increase in soluble Rho1p was observed in the cal1-1 strain grown at the restrictive temperature. The temperature-sensitive phenotype of cal1-1 is most efficiently suppressed by overproduction of Rho1p. Overproduction of the other essential target, Cdc42p, in contrast, is deleterious in cal1-1 cells, but not in other cdc43 mutants or the wild-type strains. The cdc43-5 mutant cells accumulate Cdc42p in soluble pools and cdc43-5 is suppressed by overproduction of Cdc42p. Thus, several phenotypic differences are observed among the cal1/cdc43 mutations, possibly due to alterations in substrate specificity caused by the mutations.

Mutagenesis and biochemical analysis of recombinant yeast prenyltransferases.

The use of the S. cerevisiae protein prenyltransferases as a model system for general prenyltransferase study is justified by the similarity of mechanism, substrate specificity, and evolutionarily conserved substrates with the mammalian prenyltransferases. Genetic identification of potential structural genes involved in prenyltransferase activity can be easily confirmed with biochemical assays using recombinant enzyme reconstitution. Yeast FTase and GGTase I produced in E. coli are indistinguishable from the native proteins and can be studied without interference from contaminating cellular protein prenyltransferases. Structure-function analysis of the yeast prenyltransferase subunits is also simplified by the rapidity with which mutant enzymes can be analyzed in E. coli and their biological activity characterized in yeast defective for the particular subunit gene.

Substrate characterization of the Saccharomyces cerevisiae protein farnesyltransferase and type-I protein geranylgeranyltransferase.

The in vitro substrate preferences of recombinant S. cerevisiae protein farnesyltransferase and type-I protein geranylgeranyl-transferase were determined. Proteins ending in 16 different CaaX sequences (C = cysteine, a = aliphatic amino acid, X = variable amino acids) were used to determine the protein substrate preferences of these S. cerevisiae prenyltransferases. The identities of the attached prenyl groups were confirmed by iodomethane treatment of prenylated substrates and reverse-phase HPLC. The CaaX preference of each recombinant yeast enzyme was found to be nearly identical to the reported preferences of purified mammalian protein farnesyltransferase and type-I protein geranylgeranyltransferase. S. cerevisiae farnesyltransferase preferentially farnesylated CaaX sequences ending in methionine, cysteine or serine. The farnesyltransferase also attached geranylgeranyl to some CaaX sequences ending in methionine, leucine and cysteine. The S. cerevisiae type-I geranylgeranyltransferase preferentially geranylgeranylated CaaX sequences ending in leucine and to a lesser degree methionine. Yeast extracts were found to contain prenylating activities identical to those observed with the recombinant enzymes. Farnesyltransferase activity in yeast extracts exceeded type-I geranylgeranyltransferase activity by at least 3-fold, resulting in prenylation of leucine-ending CaaX substrates with a mixture of 75% geranylgeranyl and 25% farnesyl. These results suggest that some substrate overlap may occur between the S. cerevisiae protein farnesyltransferase and the type-I protein geranylgeranyltransferase in vivo.

CDC43 and RAM2 encode the polypeptide subunits of a yeast type I protein geranylgeranyltransferase.

The question regarding the identity of the alpha and beta subunits of the yeast type I protein geranylgeranyltransferase was explored using prokaryotic expression of candidate genes. The Saccharomyces cerevisiae CDC43 and RAM2 genes were expressed in Escherichia coli and cell extracts examined for the ability to transfer [3H]geranylgeranyl diphosphate to an appropriate CaaX protein substrate. Individual expression of each gene yielded no activity; however, co-expression of the two genes resulted in high levels of [3H] geranylgeranyl incorporation into the substrate protein Ras-Cys-Val-Val-Leu. The activity was partially purified yielding approximately 12,600 units/liter. The partially purified enzyme geranylgeranylated the Ras-Cys-Val-Val-Leu, Ras-Cys-Ala-Ile-Leu, Ras-Cys-Ile-Ile-Leu, and Ras-Cys-Thr-Ile-Leu substrates but not the Ras-Cys-Val-Leu-Ser or Ras-Ser-Val-Leu-Ser substrates. The protein geranylgeranyltransferase was highly specific for geranylgeranyl diphosphate and poorly transferred farnesyl. The recombinant enzyme was indistinguishable from the native type I geranylgeranyltransferase in yeast extracts. As has been reported for the protein farnesyltransferase, the yeast type I protein geranylgeranyltransferase is also a magnesium-requiring, zinc metalloenzyme. Interestingly, the recombinant enzyme functioned with calcium as the only divalent cation, although addition of zinc increased calcium-dependent activity 2-fold.